selecting the largest variable

[R_A_G_]

Hello SAS experts,

I have a series of variables and would like to choose the variable with the largest value.

Some background:

I have ran a program that classifies the probability of  individuals belonging to a specific class(y1-y8). I need to write a code to select the highest probability among these classes. I have attached a sample file with 10,000 examinees and their probability for belonging to each class. How can I select the variable with highest value/probability?

Thank you for any help in advance

[FriedEgg]

Not quite sure from your question what type of maximum value you are looking for?

Here are two possible answers to two possible questions...

find the by row maximum value: the max function. ie. ymax=max(of y1-y8);

find the by column maximum value: one option is proc summary. ie. proc means data=out_2 max; var y1-y8; run;

[art297]

I read your request slightly differently.  Would the following provide what you want?

proc summary data=out_2;

  var y1-y8;

  output out=need (drop=_:) mean=;

run;

proc transpose data=need out=want;

run;

proc sql;

  select *

    from want

      having col1 eq max(col1)

  ;

quit;

[Ksharp]

libname xx v9 'c:\';
data want;
 set xx.out_2;
 array _y{*} y1-y8;
 max=max(of y1-y8);
 i=whichN(max,of _y{*});
 max_var=vname(_y{i});
 drop max i;
run;

Ksharp